-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtrainer.py
executable file
·195 lines (162 loc) · 7.03 KB
/
trainer.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
from __future__ import print_function
import os.path
import torch.utils.data
from torch import nn
from torch import optim
from torch.autograd import Variable
from data_loader import get_data_loader
from utils.mnist_classifier.classify import ClassifyMNIST
from tests import Tests
from models.complex_model import VAE
from models.complex_model_2 import VAE2
from models.simple_model import VAE as TargetModel
from models.discriminator import Discriminator
from utils.graph import Graph
from utils.loss import complex_loss_function as source_loss
from utils.loss import simple_loss_function as target_loss
from options import load_arguments
args = load_arguments()
args.cuda = not args.no_cuda and torch.cuda.is_available()
graph = Graph(args.graph_name)
torch.manual_seed(args.seed)
classifyMNIST = ClassifyMNIST(args)
if args.cuda:
torch.cuda.manual_seed(args.seed)
if args.source == 'mnist':
train_loader_source = get_data_loader(args, True, 'mnist')
train_loader_target = get_data_loader(args, True, 'fashionMnist')
elif args.source == 'fashionMnist':
print('data loader fashion mnist')
train_loader_source = get_data_loader(args, True, 'mnist')
train_loader_target = get_data_loader(args, True, 'mnist')
else:
raise Exception('args.source does not defined')
if not (args.resume and os.path.isfile(args.model_target_path)):
print('Creating new target model')
model_target = VAE()
else:
print('Loading target model from {}'.format(args.model_target_path))
model_target = torch.load(args.model_target_path)
if not (args.resume and os.path.isfile(args.model_source_path)):
print('Creating new source model')
model_source = VAE2()
else:
print('Loading source model from {}'.format(args.model_source_path))
model_source = torch.load(args.model_source_path)
discriminator_model = Discriminator(20, 20)
if args.cuda:
model_target.cuda()
model_source.cuda()
discriminator_model.cuda()
# target_optimizer_encoder_params = [{'params': model_target.fc1.parameters()}, {'params': model_target.fc2.parameters()}]
target_optimizer = optim.Adam(model_target.parameters(), lr=args.lr)
# target_optimizer_encoder = optim.Adam(target_optimizer_encoder_params, lr=args.lr)
source_optimizer = optim.Adam(model_source.parameters(), lr=args.lr)
d_optimizer = optim.Adam(discriminator_model.parameters(), lr=args.lr)
criterion = nn.BCELoss()
if args.source == 'mnist':
tests = Tests(model_source, model_target, classifyMNIST, 'mnist', 'fashionMnist', args, graph)
elif args.source == 'fashionMnist':
tests = Tests(model_source, model_target, classifyMNIST, 'fashionMnist', 'mnist', args, graph)
else:
raise Exception('args.source does not defined')
def reset_grads():
model_target.zero_grad()
model_source.zero_grad()
discriminator_model.zero_grad()
def gen(model, input_data, optimizer, loss, batch):
reset_grads()
decode, mu, logvar, _ = model(input_data)
loss_generator = loss(decode, input_data, mu, logvar, batch)
loss_generator.backward()
optimizer.step()
running_counter = 0
for epoch in range(1, args.epochs + 1):
print('---- Epoch {} ----'.format(epoch))
iteration = 0
model_source.train()
model_target.train()
discriminator_model.train()
source_iter = iter(train_loader_source)
target_iter = iter(train_loader_target)
# ---------- Train --------------
while iteration < len(source_iter) and iteration < len(target_iter):
running_counter += 1
iteration += 1
source_input, _ = source_iter.next()
source_input = Variable(source_input)
target_input, _ = target_iter.next()
target_input = Variable(target_input)
if args.cuda:
source_input = source_input.cuda()
target_input = target_input.cuda()
# Train generators
gen(model_source, source_input, source_optimizer, source_loss, args.batch_size)
gen(model_target, target_input, target_optimizer, source_loss, args.batch_size)
# reset_grads()
# decode_t, mu_t, logvar_t, z_t = model_target(target_input)
# t_loss_generator = source_loss(decode_t, target_input, mu_t, logvar_t, args)
# t_loss_generator.backward()
# target_optimizer.step()
# reset_grads()
# decode_s, mu_s, logvar_s, z_s = model_source(source_input)
# s_loss_generator = source_loss(decode_s, source_input, mu_s, logvar_s, args)
# s_loss_generator.backward()
# source_optimizer.step()
# Train encoder
# d_fake_t = discriminator_model(z_t)[:, 0]
# t_loss_discriminator = criterion(d_fake_t, ones)
# if args.one_sided:
# t_loss = t_loss_discriminator
# else:
# t_loss = args.h_tg * t_loss_generator + t_loss_discriminator
#
# d_fake_s = discriminator_model(z_s)[:, 0]
# s_loss_discriminator = criterion(d_fake_s, ones)
# if args.apply_source_to_discriminator:
# s_loss = s_loss_generator + s_loss_discriminator
# else:
# s_loss = s_loss_generator
# reset_grads()
# Train Discriminator
# z_s = z_s.detach()
# _, _, _, z_s = model_source(source_input)
# d_real_decision = discriminator_model(z_s)[:, 0]
# d_real_error = criterion(d_real_decision, ones) # ones = true
# d_real_error.backward()
#
# # z_t = z_t.detach()
# _, _, _, z_t = model_target(target_input)
# d_fake_decision = discriminator_model(z_t)[:, 0]
# d_fake_error = criterion(d_fake_decision, zeros) # zeros = fake
# d_fake_error.backward()
# d_optimizer.step()
# for p in discriminator_model.parameters():
# p.data.clamp_(-0.1, 0.1)
graph.last1 = 0#s_loss_generator.data[0]
graph.last2 = 0#t_loss_generator.data[0]
graph.last3 = 0#s_loss_discriminator.data[0]
graph.last4 = 0#t_loss_discriminator.data[0]
graph.last5 = 0#d_real_error.data[0]
graph.last6 = 0#d_fake_error.data[0]
graph.add_point(running_counter)
# ---------- Tests --------------
# tests.source_to_target_test()
# tests.args.one_sided = not tests.args.one_sided
# tests.source_to_target_test()
# tests.args.one_sided = not tests.args.one_sided
# tests.gaussian_input()
tests.tsne(train_loader_source, model_source)
tests.tsne(train_loader_target, model_target)
# if not args.one_sided:
# tests.reconstruction(epoch)
# certain, sparse = tests.test_matching()
# accuracy = certain + sparse
# print('certain: {}, sparse: {}, all: {} old max: {}'.format(certain, sparse, accuracy, overall_accuracy))
# if epoch > 10 and accuracy > overall_accuracy:
# overall_accuracy = accuracy
# print('saving mnist model')
# if not os.path.isdir('results/{}'.format(args.graph_name)):
# os.mkdir('results/{}'.format(args.graph_name))
# torch.save(model_source, 'results/{}/model_source.pt'.format(args.graph_name))
# torch.save(model_target, 'results/{}/model_target.pt'.format(args.graph_name))